Preclinical positron emission tomography (PET) has become a widely used tool in biomedical research, particularly in the evaluation of new therapeutics. Hundreds of scanners are installed across the world in major medical research centers and within pharmaceutical companies. For a variety of reasons, however, the performance of these systems falls well short of what can theoretically be achieved. This has two important consequences. Firstly, the quantitative potential of current studies of radiopharmaceutical kinetics and uptake is undermined by limited spatial resolution (reducing accuracy) and limited sensitivity (reducing precision). Secondly, important applications for preclinical PET, for example metabolic imaging in gray matter structures in the mouse brain or studies of low abundance protein targets, such as TSPO receptor expression in chronic inflammation, are just out of the reach of current instruments. The goal of this proposal is to develop a pathway towards small-animal PET scanners that can come as close as possible to the theoretical limits of spatial resolution and sensitivity imposed by fundamental physics and the properties of available detector materials. Building on 15 years of development of increasingly high resolution detectors for small-animal PET, recent advances in realizing dual-ended detectors that can also simultaneously provide high sensitivity and combining these with highly innovative silicon photomultiplier (SiPM) photodetectors, we propose a design that will lead to detector modules with unprecedented performance for small-animal PET applications. Specifically, we will develop fully-engineered detector modules suitable for close packing in a preclinical PET scanner geometry that will support better than 0.6 mm reconstructed spatial resolution, an average sensitivity of >10% across the whole body of a mouse, depth-of-interaction resolution < 3 mm, timing resolution < 3 ns and energy resolution < 30%. We will develop the electronics and software to efficiently read out these modules with no significant degradation in performance and integrate them into the open source OpenPET libraries for access by the entire nuclear medical imaging community. Lastly, we will use experimental data from fully-engineered detector modules combined with advanced simulation tools to design and predict the performance of a preclinical scanner using our new technology. The outcome of this proposal will be detector modules and a scanner design that advance preclinical PET to new levels of spatial resolution and sensitivity together with the comprehensive set of experimental and simulation data that will be needed to justify moving to the next stage of developing a prototype small-animal PET scanner. There also will be a broader impact in that the detector technology and electronics developed will be applicable to other PET systems, especially dedicated organ imaging (e.g. brain, breast) scanners and PET/MR systems.